Development of a protective HIV vaccine would have great impact in stemming the HIV pandemic. Placebo-controlled preventative HIV vaccine efficacy trials and controlled non-human primate (NHP) vaccine challenge trials are important components of vaccine development. This renewal application proposes to continue development of novel statistical methods to address key issues in the design and analysis of such trials. Aim 1 (Sieve Analysis) develops statistical methods for evaluating the impact of HIV genotypic and phenotypic variation on the effects of HIV vaccines evaluated in efficacy trials. The methods build on previous work to address three key challenges: (1) ideally the sieve analysis assesses HIV sequences measured on acute-phase samples of breakthrough infected subjects, yet for some subjects only post-acute-phase sequences can be measured; (2) the analysis of longitudinal HIV sequences is needed to discriminate whether an observed difference in sequences (vaccine-vs-placebo) is due to strain-specific blockage of HIV acquisition or to post-acquisition T-cell driven HIV sequence evolution; and (3) achieving high statistical power requires that the measure of genetic distance between a subject's breakthrough sequences and the vaccine-insert sequence be highly immunologically relevant, based on T cell epitopes that a subject can recognize given his/her HLA alleles. Aim 2 (Immune Correlates) develops statistical methods for evaluating vaccine-induced immune responses as surrogates of protection within vaccine efficacy trials. The methods build on previous work for evaluating an immunological surrogate endpoint, with new directions to: evaluate and compare surrogate value of multiple immunological biomarkers; improve efficiency by leveraging baseline covariates; improve robustness by weakening assumptions; develop an innovative one-way crossover design; provide assumption-diagnostics and sensitivity analyses; and develop meta-analysis methods for evaluating immunological surrogates over multiple efficacy trials. The methods will be applied to the six HIV vaccine efficacy trials that have been conducted. The meta-analysis methods will also be employed to predict efficacy of influenza A H1N1 vaccines. Aim 3 (NHP Trials) develops statistical methods for evaluating vaccine efficacy on SIV acquisition, viral load, and survival endpoints, and for evaluating immunological surrogates for these endpoints within NHP trials. The methods (1) provide exact estimation and testing of vaccine efficacy, leveraging baseline covariates to improve efficiency; (2) address the challenge that viral load trajectories are truncated by death; and (3) utilize novel study designs for evaluating surrogate endpoints. This research will be developed in collaboration with international leaders in pre-clinical HIV vaccine evaluation, and will be applied to several large NHP repeated low-dose challenge trials.